Re: [opensuse] hda -> sda
On 06/10/31 14:43 (GMT+0100) Christoph Thiel apparently typed:
On Tue, Oct 31, 2006 at 07:23:11AM -0500, Felix Miata wrote:
Check out this too: https://bugzilla.novell.com/show_bug.cgi?id=214992
We'll be using classical ide drivers for PATA devices in Beta2 -- so this bug is workarounded.
The bigger problem is as described in https://bugzilla.novell.com/show_bug.cgi?id=214992#c3 which says it's a kernel problem that I have been able to confirm in http://bugzilla.kernel.org/. I can't put any Linux on this system because the target partition is non-negotiably sata #16 or higher. If I want to install linux anywhere I choose on a disk, I'm stuck with PATA. Windoz makes no such imposition on how I logically carve up my huge disks. -- "The Lord is my strength and my shield; my heart trusts in him, and I am helped." Psalm 28:7 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Felix Miata wrote:
On 06/10/31 14:43 (GMT+0100) Christoph Thiel apparently typed:
On Tue, Oct 31, 2006 at 07:23:11AM -0500, Felix Miata wrote:
Check out this too: https://bugzilla.novell.com/show_bug.cgi?id=214992
We'll be using classical ide drivers for PATA devices in Beta2 -- so this bug is workarounded.
The bigger problem is as described in https://bugzilla.novell.com/show_bug.cgi?id=214992#c3 which says it's a kernel problem that I have been able to confirm in http://bugzilla.kernel.org/.
I can't put any Linux on this system because the target partition is non-negotiably sata #16 or higher. If I want to install linux anywhere I choose on a disk, I'm stuck with PATA. Windoz makes no such imposition on how I logically carve up my huge disks.
I had a patch to use up to 256 partitions per disk independent of the type of disk. I can dig it out if anyone wants to test. Regards, Carl-Daniel -- http://www.hailfinger.org/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tue, 31 Oct 2006, Carl-Daniel Hailfinger wrote:
Felix Miata wrote:
On 06/10/31 14:43 (GMT+0100) Christoph Thiel apparently typed:
On Tue, Oct 31, 2006 at 07:23:11AM -0500, Felix Miata wrote:
Check out this too: https://bugzilla.novell.com/show_bug.cgi?id=214992 We'll be using classical ide drivers for PATA devices in Beta2 -- so this bug is workarounded. The bigger problem is as described in https://bugzilla.novell.com/show_bug.cgi?id=214992#c3 which says it's a kernel problem that I have been able to confirm in http://bugzilla.kernel.org/.
I had a patch to use up to 256 partitions per disk independent of the type of disk. I can dig it out if anyone wants to test.
Could I get your patch? I think it would be good if this were readily available to all. Thanks, -- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047 --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
Boyd Lynn Gerber wrote:
On Tue, 31 Oct 2006, Carl-Daniel Hailfinger wrote:
Felix Miata wrote:
On 06/10/31 14:43 (GMT+0100) Christoph Thiel apparently typed:
On Tue, Oct 31, 2006 at 07:23:11AM -0500, Felix Miata wrote:
Check out this too: https://bugzilla.novell.com/show_bug.cgi?id=214992 We'll be using classical ide drivers for PATA devices in Beta2 -- so this bug is workarounded. The bigger problem is as described in https://bugzilla.novell.com/show_bug.cgi?id=214992#c3 which says it's a kernel problem that I have been able to confirm in http://bugzilla.kernel.org/. I had a patch to use up to 256 partitions per disk independent of the type of disk. I can dig it out if anyone wants to test.
Could I get your patch? I think it would be good if this were readily available to all.
Ahem... I forgot that you need a recompiled kernel for it to work. Or at least arcane and forgotten commands like partx. Personally, I like the partx way better. With the attached patch to util-linux you can use partx. Make sure no partition on sda is in use (swap/suspend/mounted), then # partx -d /dev/sda # kpartx -a /dev/sda Now /dev/sda[1-15] have vanished and /dev/disk/by-name/sda[1-255] appear instead. Mounting these devices works normally. I also added this to https://bugzilla.novell.com/show_bug.cgi?id=214992#c12 Regards, Carl-Daniel -- http://www.hailfinger.org/ --- util-linux-2.12r/Makefile 2004-12-23 01:43:09.000000000 +0100 +++ util-linux-2.12r/Makefile 2006-11-02 16:44:52.000000000 +0100 @@ -18,7 +18,8 @@ fdisk \ hwclock \ sys-utils \ - text-utils + text-utils \ + partx .PHONEY: all install clean now all: defines.h @@ -43,7 +44,7 @@ clean: -rm -f *.o *~ omake conftest conftest.c core - @for subdir in $(SUBDIRS) partx; do \ + @for subdir in $(SUBDIRS); do \ (cd $$subdir && $(MAKE) $@) || exit 1; \ done --- util-linux-2.12r/partx/partx.c~ 2004-08-23 22:13:27.000000000 +0200 +++ util-linux-2.12r/partx/partx.c 2006-11-02 17:05:55.000000000 +0100 @@ -38,6 +38,8 @@ #include <unistd.h> #include <sys/ioctl.h> #include <linux/hdreg.h> /* HDIO_GETGEO */ +/* workaround for __user in header file */ +#define __user #include <linux/blkpg.h> #define BLKGETSIZE _IO(0x12,96) /* return device size */ --- util-linux-2.12r/partx/addpart.c~ 2006-11-02 17:18:06.000000000 +0100 +++ util-linux-2.12r/partx/addpart.c 2006-11-02 17:18:46.000000000 +0100 @@ -3,6 +3,8 @@ #include <fcntl.h> #include <stdlib.h> #include <sys/ioctl.h> +/* workaround for __user in header file */ +#define __user #include <linux/blkpg.h> int --- util-linux-2.12r/partx/delpart.c~ 1999-07-09 04:56:36.000000000 +0200 +++ util-linux-2.12r/partx/delpart.c 2006-11-02 17:18:52.000000000 +0100 @@ -3,6 +3,8 @@ #include <fcntl.h> #include <stdlib.h> #include <sys/ioctl.h> +/* workaround for __user in header file */ +#define __user #include <linux/blkpg.h> int --- util-linux-2.12r/partx/Makefile~ 2002-04-10 12:11:07.000000000 +0200 +++ util-linux-2.12r/partx/Makefile 2006-11-02 17:33:53.000000000 +0100 @@ -1,7 +1,12 @@ +include ../MCONFIG + OBJ = bsd.o dos.o partx.o solaris.o unixware.o gpt.o crc32.o CFLAGS += -Wall +SBIN= + +SBIN:=$(SBIN) partx -all: addpart delpart partx +all: addpart delpart $(SBIN) partx: $(OBJ) @@ -13,3 +18,7 @@ rm -f $(OBJ) addpart.o delpart.o addpart delpart partx *~ $(OBJ): partx.h + +install: all + $(INSTALLDIR) $(SBINDIR) + $(INSTALLBIN) $(SBIN) $(SBINDIR)
Hi, On Tue, 31 Oct 2006, Felix Miata wrote:
On 06/10/31 14:43 (GMT+0100) Christoph Thiel apparently typed:
On Tue, Oct 31, 2006 at 07:23:11AM -0500, Felix Miata wrote:
Check out this too: https://bugzilla.novell.com/show_bug.cgi?id=214992
We'll be using classical ide drivers for PATA devices in Beta2 -- so this bug is workarounded.
The bigger problem is as described in https://bugzilla.novell.com/show_bug.cgi?id=214992#c3 which says it's a kernel problem that I have been able to confirm in http://bugzilla.kernel.org/.
I can't put any Linux on this system because the target partition is non-negotiably sata #16 or higher. If I want to install linux anywhere I choose on a disk, I'm stuck with PATA. Windoz makes no such imposition on how I logically carve up my huge disks.
Having a maximum of 15 partitions an a SCSI disk is not a linux kernel problem, it is simply a linux kernel fact. Cheers -e -- Eberhard Moenkeberg (emoenke@gwdg.de, em@kki.org) --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 06/10/31 09:54 (GMT-0500) Eberhard Moenkeberg apparently typed:
On Tue, 31 Oct 2006, Felix Miata wrote:
I can't put any Linux on this system because the target partition is non-negotiably sata #16 or higher. If I want to install linux anywhere I choose on a disk, I'm stuck with PATA. Windoz makes no such imposition on how I logically carve up my huge disks.
Having a maximum of 15 partitions an a SCSI disk is not a linux kernel problem, it is simply a linux kernel fact.
Exact same difference, except that the limitation when SCSI was SCSI was SCSI and not SCSI/USB/SATA/Firewire/whoknowswhatelse had a minor universe of people it affected. With the replacement of legacy ata by libata (pretending SATA is SCSI), it morphed into an artificial limitation that potentially impacts a universe several magnitudes larger. That means a solution is called for, no matter what the kernel developers say or do or not about it, or how the status quo is defined. I don't really know anything about kernel development, but if it was up to me, with the 2.8 generation of kernels, the majors and minors for PATA & SATA would be swapped to permit max 63 on SCSI and instead limit PATA to 15 unless a superior integrated system could be found and implemented. My disks average upwards of 20 partitions each, a number which continues to climb. This 24/7 system has two disks, with 43 and 18 partitions. My other 24/7 system has 1 disk with 22 partitions. There's simply no way for those numbers to be reduced as long as disk sizes keep escalating. -- "The Lord is my strength and my shield; my heart trusts in him, and I am helped." Psalm 28:7 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 31 October 2006 2:36 pm, Felix Miata wrote:
My disks average upwards of 20 partitions each, a number which continues to climb. This 24/7 system has two disks, with 43 and 18 partitions. My other 24/7 system has 1 disk with 22 partitions. There's simply no way for those numbers to be reduced as long as disk sizes keep escalating.
If you need so many partitions, why are you not using LVM? -- James Oakley jfunk@funktronics.ca --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 06/10/31 14:39 (GMT-0400) James Oakley apparently typed:
On Tuesday 31 October 2006 2:36 pm, Felix Miata wrote:
My disks average upwards of 20 partitions each, a number which continues to climb. This 24/7 system has two disks, with 43 and 18 partitions. My other 24/7 system has 1 disk with 22 partitions. There's simply no way for those numbers to be reduced as long as disk sizes keep escalating.
If you need so many partitions, why are you not using LVM?
I have no knowledge of any way LVM could do anything but increase complication. My backup strategy is heavily dependent on cloning disks and partitions across 20 machines as much as 8-10 years old having various multiboot mixtures of OS/2, DOS, doz, and Linux. -- "The Lord is my strength and my shield; my heart trusts in him, and I am helped." Psalm 28:7 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Tuesday 31 October 2006 3:18 pm, Felix Miata wrote:
On 06/10/31 14:39 (GMT-0400) James Oakley apparently typed:
On Tuesday 31 October 2006 2:36 pm, Felix Miata wrote:
My disks average upwards of 20 partitions each, a number which continues to climb. This 24/7 system has two disks, with 43 and 18 partitions. My other 24/7 system has 1 disk with 22 partitions. There's simply no way for those numbers to be reduced as long as disk sizes keep escalating.
If you need so many partitions, why are you not using LVM?
I have no knowledge of any way LVM could do anything but increase complication. My backup strategy is heavily dependent on cloning disks and partitions across 20 machines as much as 8-10 years old having various multiboot mixtures of OS/2, DOS, doz, and Linux.
Even more reason for LVM in addition to Xen/Qemu or VMware. You can consolidate your old machines into a couple of good ones and do your full disk backups with minimal downtime using the LVM tools (no downtime in some cases). That approach also allows you to light up your virtual machines immediately on any of your boxes in case one goes down. -- James Oakley jfunk@funktronics.ca --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On 06/10/31 15:04 (GMT-0500) James Oakley apparently typed:
On Tuesday 31 October 2006 3:18 pm, Felix Miata wrote:
On 06/10/31 14:39 (GMT-0400) James Oakley apparently typed:
On Tuesday 31 October 2006 2:36 pm, Felix Miata wrote:
My disks average upwards of 20 partitions each, a number which continues to climb. This 24/7 system has two disks, with 43 and 18 partitions. My other 24/7 system has 1 disk with 22 partitions. There's simply no way for those numbers to be reduced as long as disk sizes keep escalating.
If you need so many partitions, why are you not using LVM?
I have no knowledge of any way LVM could do anything but increase complication. My backup strategy is heavily dependent on cloning disks and partitions across 20 machines as much as 8-10 years old having various multiboot mixtures of OS/2, DOS, doz, and Linux.
Even more reason for LVM in addition to Xen/Qemu or VMware.
It's less reason, not more reason.
You can consolidate your old machines into a couple of good ones and do your full disk backups with minimal downtime using the LVM tools (no downtime in some cases). That approach also allows you to light up your virtual machines immediately on any of your boxes in case one goes down.
It's an enticing idea, but it's unworkable, and it's atithetical to the object of many copies spread among the many failure prone devices. HDs are hopelessly unreliable in the long run, and yet the best we have available. It's unworkable because a reason, among others, for having many systems, is to be able to reproduce as many hardware/OS permutations as possible. Then there's the additional complication that only one of them has sufficient RAM to support an insanely RAM hungry virtual machine environment. Lastly, I see LVM as additional complication, not simplification. Life with multiple systems is complicated enough without the addition of that extra layer, or concentrating more dependance on fewer failure prone physical devices. -- "The Lord is my strength and my shield; my heart trusts in him, and I am helped." Psalm 28:7 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
I don't really know anything about kernel development, but if it was up to me, with the 2.8 generation of kernels, the majors and minors for PATA & SATA would be swapped to permit max 63 on SCSI and instead limit PATA to 15 unless a superior integrated system could be found and implemented.
It's not like we don't have enough majors/minors. :) -`J' -- --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (6)
-
Boyd Lynn Gerber -
Carl-Daniel Hailfinger -
Eberhard Moenkeberg -
Felix Miata -
James Oakley -
Jan Engelhardt